1554208 - Implement FT2 backend for the cross-process shared font list

Status: RESOLVED FIXED

(Core :: Layout: Text and Fonts, enhancement, P3)

Description

[Jonathan Kew [:jfkthame]]

In bug 1514869, an initial implementation of the shared font list landed for desktop platforms. This bug is for adding support to gfxFT2FontList, so that it will be available in Fenix as well.

Comment 1

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1554208 - patch 1 - Add missing checks for shared charmap. r=jwatt

This isn't really specific to the FT2 fontlist, it's a general fixup that I noticed while
working on this. (The missing checks aren't crucial, but mean that we might re-read a cmap
when we shouldn't need to.)

Comment 2

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1554208 - patch 2 - Remove redundant member-reset code from gfxFT2FontList::FindFonts (already handled by gfxPlatformFontList before this is called). r=jwatt

A bit of cleanup of the existing code, before we start actually implementing new stuff.

Depends on D36106

Comment 3

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1554208 - patch 3 - Rearrange gfxFT2Fonts code so that FindFonts() does just what it says, and other work is handled by the caller. r=jwatt

This makes the functional structure a bit cleaner, so that it'll be easier to slip in the alternative
codepath for the shared font-list.

Depends on D36107

Comment 4

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1554208 - patch 4 - Make the FontNameCache that we use to store the font list in the startupCache use a more robust and extensible serialization. r=jwatt

I'll need to add a couple of extra fields to the cache records, and realized that the current format
looks really fragile; in theory, it'd break if a font name ever contained a comma or semicolon
(unlikely though that may be). So let's fix it to be a bit more robust before we add to it further.
Using control characters from the C0 range to delimit fields/records, instead of ASCII punctuation,
removes the (tiny) risk of conflict with characters that actually occur in a name, and using
distinct field and record separators means that we can better check that the records we're loading
from the cache actually match the expected format.

(Given that the startup cache gets re-created when the build ID is updated, a change in the format
between versions wasn't going to cause problems for users; their old cache just gets blown away
when upgrading. Still, a little more robustness seems like a good thing.)

Depends on D36108

Comment 5

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1554208 - patch 5 - Implement shared-fontlist support in the gfxFT2FontList backend. r=jwatt

This is the main part of the implementation, except that it doesn't handle populating the
local names table (for @font-face src:local() lookups) with Full and PostScript names;
that follows in the next patch.

Depends on D36109

Comment 6

[Jonathan Kew [:jfkthame]]

Attachment: Bug 1554208 - patch 6 - Add PSName and FullName support to gfxFT2FontList shared font-list initialization, so that src:local() lookups will work. r=jwatt

And with this, all tests pass on tryserver when the shared list is enabled.

Depends on D36110

Comment 7

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/b610a6988076
patch 1 - Add missing checks for shared charmap. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/3e824f3e3c9c
patch 2 - Remove redundant member-reset code from gfxFT2FontList::FindFonts (already handled by gfxPlatformFontList before this is called). r=jwatt
https://hg.mozilla.org/integration/autoland/rev/f72e750ec20d
patch 3 - Rearrange gfxFT2Fonts code so that FindFonts() does just what it says, and other work is handled by the caller. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/3e6c11ef906c
patch 4 - Make the FontNameCache that we use to store the font list in the startupCache use a more robust and extensible serialization. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/7bd8be02631a
patch 5 - Implement shared-fontlist support in the gfxFT2FontList backend. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/14b2c3daeaf0
patch 6 - Add PSName and FullName support to gfxFT2FontList shared font-list initialization, so that src:local() lookups will work. r=jwatt

Comment 8

[Cristian Brindusan [:cbrindusan]]

Backed out 6 changesets (Bug 1554208) for reftest failures at font-sans-serif.html and geckoview failures at AccessibilityTest.testMoveByLine.

Backout: https://hg.mozilla.org/integration/autoland/rev/c91bf14c149c2fbb986910ffb479248151138f53

Push that started the failures: https://treeherder.mozilla.org/#/jobs?repo=autoland&resultStatus=pending%2Crunning%2Csuccess%2Ctestfailed%2Cbusted%2Cexception&revision=14b2c3daeaf0d4e41fc9988ca1f62a711f7ad710&selectedJob=255327572

Failure log:
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=255327572&repo=autoland&lineNumber=2009
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=255326521&repo=autoland&lineNumber=1477

Comment 9

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/58dce111eb0c
patch 1 - Add missing checks for shared charmap. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/54c09f3cf6d6
patch 2 - Remove redundant member-reset code from gfxFT2FontList::FindFonts (already handled by gfxPlatformFontList before this is called). r=jwatt
https://hg.mozilla.org/integration/autoland/rev/0efb6fa15fb7
patch 3 - Rearrange gfxFT2Fonts code so that FindFonts() does just what it says, and other work is handled by the caller. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/bb4682a23c79
patch 4 - Make the FontNameCache that we use to store the font list in the startupCache use a more robust and extensible serialization. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/8453b0c76891
patch 5 - Implement shared-fontlist support in the gfxFT2FontList backend. r=jwatt
https://hg.mozilla.org/integration/autoland/rev/3261c2a5d165
patch 6 - Add PSName and FullName support to gfxFT2FontList shared font-list initialization, so that src:local() lookups will work. r=jwatt

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/ea16d55620de
update test expectations for unexpected passes on Android. a=unexpected-pass-fixes

Comment 11

[Jonathan Kew [:jfkthame]]

(In reply to Pulsebot from comment #10)

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/ea16d55620de
update test expectations for unexpected passes on Android.
a=unexpected-pass-fixes

Thanks for this! So try with "-u all" doesn't run web-platform-tests on Android, huh?

(I assume the reason for the new passes here is that we now properly maintain the original form of the font-family name as declared in CSS, and hence "Ahem" now matches our special-case code for disabling antialiasing during these tests; previously it was getting lowercased and slipping through.)

Comment 12

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/3e793ca066f2
Revert expectation for font-matching-rule-009.xht because of permafails. a=testonly

Comment 13

[Cosmin Sabou [:CosminS]]

https://hg.mozilla.org/mozilla-central/rev/58dce111eb0c
https://hg.mozilla.org/mozilla-central/rev/54c09f3cf6d6
https://hg.mozilla.org/mozilla-central/rev/0efb6fa15fb7
https://hg.mozilla.org/mozilla-central/rev/bb4682a23c79
https://hg.mozilla.org/mozilla-central/rev/8453b0c76891
https://hg.mozilla.org/mozilla-central/rev/3261c2a5d165
https://hg.mozilla.org/mozilla-central/rev/ea16d55620de
https://hg.mozilla.org/mozilla-central/rev/3e793ca066f2